Genotype-by-Environment Interaction for Quantitative Trait Loci Affecting Nitrogen Use Efficiency and Associated Traits in Potato


Deciphering the genetic basis of complex traits like nitrogen use efficiency (NUE) requires understanding the genotype-by-environment (G × E) interaction and linking physiological functions and agronomic traits to DNA markers. Multi-environment experiments were conducted in different environments representing low and high nitrogen levels combined with rain-fed and irrigation production conditions at three different locations in Ethiopia: Debre-Tabor, Injibara and Koga, in 2013 and 2015. The objectives of the study were to determine the G × E interaction and stability of genotypes for NUE of potato and to identify markers associated with NUE and NUE-related agronomic and physiological traits in potato under these diverse environments. Data were analyzed using GenStat, and genotype plus the genotype and environment (GGE) biplot model; the marker-trait associations were discovered using the R-software package GWASpoly. The analysis of variance that included location and production system had estimates for genotype variance (σ2g) that were low compared with the estimates for environment variance (σ2e) and genotype-by-environment interaction variance (σ2ge) for most measured traits at both N levels. The GGE analysis identified two mega-environments that coincided with the two production systems. The high N level environments both at Debre-Tabor and Injibara, and the low N environment at Koga, respectively, were the most suitable environments for discriminating the potato cultivars and being representative test environments for NUE evaluation in the rain-fed mega-environment and irrigation mega-environment. A total of 77 marker trait associations were identified for NUE and NUE-related agronomic and physiological traits. Multi-trait genomic regions that harbored significant marker-trait associations for NUE and NUE-related traits were found on chromosomes III, V and VI. The effect of production season was greater than the effect of N levels on QTL × environment interaction for most NUE-related traits.